Adaptive landscapes of transcription factors and their in vivo binding sites in 1135 Arabidopsis thaliana genomes
Bioinformatics and Theoretical Biology
Final Report Abstract
Regulated gene expression is essential for building and maintaining organismal phenotypes. Gene expression is controlled via several molecular mechanisms, including the binding of a transcription factor to specific sequence on DNA. In this work, I investigated how transcription factor binding sequences evolve in natural populations of Arabidopsis thaliana. To this end, I integrated results from previously published work. These data included (i) data on in vitro binding affinities of a transcription factor to double-stranded DNA, as determined through protein binding microarray experiments; (ii) data on genomic binding regions of transcription factors identified through in vitro DNA affinity purification and sequencing; (iii) genomic footprint data from a DNase I hypersensitivity in vivo experiment with root tissue from A. thaliana, and (iv) genomic polymorphism data from a worldwide collection of 1,135 A. thaliana accessions. I used the results to first identify bound genomic loci in the reference accession Col-0, and to then obtain orthologous binding sequences in additional 1,134 accessions. Moreover, results from protein binding microarrays allowed me to link each binding sequence with quantitative measurements of binding affinities to a transcription factor. With the help of these resources, I studied 19 different transcription factors and identified 8,333 genomic loci bound in vivo by these transcription factors. Using these data, I explored factors that may contribute to population frequencies of binding sequences. First, I investigated if more frequent binding sequences are associated with stronger binding affinities. This is indeed the case for all 19 analyzed transcription factors. Second, I hypothesized that a binding sequence may be frequent in a population, because it can be created via multiple accessible mutational paths. Each sequence in a population can be created from every other sequence in the sequence space (all possible nucleotide sequences) through a succession of mutations, and I call a path accessible if no mutation decreases binding affinities along this path. I again found that more frequent binding sequences can be created by a larger number of accessible mutational paths. Epistatic interactions among individual mutation pairs that alter binding affinity are pervasive and can help explain variation in accessibility among binding sequences. However, I did not find evidence that the presence of epistasis itself is evolving under selection. Using partial correlation analysis, I showed that both binding affinity and path accessibility contribute to allele frequencies of binding sequences. Depending on the transcription factor, one of the two variables was found to be more important than the other. In summary, I show how the combination of in vitro binding affinity data with in vivo binding sequence data can help understand the forces that affect the evolution of transcription factor binding sequences in natural populations. In particular, I demonstrate that high fitness is not always the only explanation for the prevalence of frequent binding sequences in A. thaliana.
Publications
- (2020) Genotype networks of 80 quantitative Arabidopsis thaliana phenotypes reveal phenotypic evolvability despite pervasive epistasis. PLoS Comput Biol 16(8): e1008082
Schweizer G, Wagner A
(See online at https://doi.org/10.1371/journal.pcbi.1008082) - (2021) Both Binding Strength and Evolutionary Accessibility Affect the Population Frequency of Transcription Factor Binding Sequences in Arabidopsis thaliana. Genome Biol Evol 13(12): evab273
Schweizer G, Wagner A
(See online at https://doi.org/10.1093/gbe/evab273)
